Background
Tuberculosis (TB) remains one of the major causes of morbidity and mortality worldwide with the highest burden found in Africa and Asia, mainly linked to the human immunodeficiency virus (HIV) epidemic [
1]. The 2016 World Health Organization (WHO) report on TB revealed that there were an estimated 10 million new cases of TB in the year 2015 worldwide, with 11% having HIV. An estimated 1.8 million people died due to TB in 2015, including 0.4 million deaths due to HIV/TB co-infection. The incidence of TB among HIV patients in Uganda was estimated at 202 new cases per 100,000 population [
2].
Due to the limited published data on fungal disease epidemiology in sub-Saharan Africa, a recent review attempted to estimate the burden of fungal infections in Uganda using specific populations [
3]. In this review, chronic pulmonary aspergillosis (CPA) was estimated at 12–22% in TB patients with cavities and 1–4% in those without cavities. Considering post-TB data in Uganda, asymptomatic CPA was estimated at 7% with an additional 1.7% having detectable
Aspergillus-specific immunoglobulin G (IgG) antibodies with cavitation.
Recent work done in Northern Uganda has validated some of these estimates, with a CPA prevalence of 8.2%, and 6.7% having cavities among patients who had been successfully treated for pulmonary TB within the last 7 years [
4,
5]. More results from this work showed
Aspergillus-specific IgG antibody levels were raised in 26% of patients with “smear negative TB” and suggested that previously unrecognized CPA might be responsible for significant mortality in patients treated for TB in Uganda [
6‐
8]. Beyond this limited data, little is known about the epidemiology of fungal colonisation and sensitisation, and their contribution to TB disease progress and treatment outcomes in Uganda where pulmonary TB is very common, in part driven by the high prevalence of HIV [
9].
We hypothesized that patients with pulmonary TB may get colonized with
Aspergillus during and in the post treatment period leading to a chronic lung infection and/or allergic fungal disease if the patient was pre-sensitized to
Aspergillus antigens. Pulmonary cavitation is a pre-disposing factor for CPA [
10] and may or may not present together with
Aspergillus sensitization. We therefore aimed to establish and compare
Aspergillus-specific antibody levels among HIV-infected Ugandans with TB, at the beginning and end of TB treatment; using ImmunoCAP® and Immulite® immunoassays.
Discussion
The study demonstrated that
Aspergillus-specific IgG antibodies were elevated in 4% of HIV-infected Ugandan adults at the start of TB treatment and in 9% at the end of TB treatment. Using ImmunoCAP, participants with CD4 T-cell counts <100 cell/μl had more elevated
Aspergillus-specific IgG antibodies (
P = 0.01). The increased level of immunosuppression could have increased their susceptibility to the opportunistic fungal infection. Hyper IgG levels are common in acquired immune deficiency syndrome (AIDS) patients and represent an unrestrained B-cell response in the absence of T-cells [
14]. The ImmunoCAP
Aspergillus-specific IgG antibody titers were higher with more positives at end of TB treatment than baseline, even though the difference in means was small. However, this difference was statistically significant (
P = 0.02). Initiation of antiretroviral therapy and anti-TB medication could have improved the patients’ ability to mount a good immune response at week 24 than baseline.
Patients who were not on antiretroviral therapy at TB diagnosis (68%) had significantly higher Aspergillus-specific IgG titers (P = 0.03) using Immulite, and 45% (14/31) of them had CD4 T-cell counts less than 200 cells/μl. Increased level of immunosuppression could still be the contributing factor to this increased susceptibility to the opportunistic fungal infection. There was no significant difference in the Immulite Aspergillus-specific IgG antibody levels (P = 0.66) between baseline and end of TB treatment.
We observed a discrepancy in the
Aspergillus-specific IgG antibody levels between the ImmunoCAP and Immulite. Previous studies comparing the ImmunoCAP and Immulite have shown that the values for both tests are highly correlated with antibody levels measured by Immulite having a mean of 3 times higher than when measured by ImmunoCAP. However, this correlation is lost at high ImmunoCA
P values [
15]. This ratio may change based on the specific antibody in question. In our current study, this ratio was approximately 1.6 on average.
About 10% (9/93) of the participants had evidence of
Aspergillus sensitization at the end of TB treatment. This was an expected outcome in this population since persistence of pulmonary cavities after successful pulmonary TB treatment is very common [
16] and these cavities are thought to harbour mould spores leading to fungal colonisation. Besides, ABPA can be misdiagnosed as pulmonary TB, with some similar clinical features [
10,
17]. However, ABPA is rarely described outside chronic obstructive pulmonary disease (COPD), asthma and cystic fibrosis. There was no record of these three conditions in our participants, and therefore we argue that
Aspergillus sensitisation is the explanation for these high IgE results, which could represent a strong T-helper 2 (Th2) response. The association between ABPA and pulmonary TB has been weakly described before mainly in case reports [
18‐
20].
In a recent review, CPA in Uganda was estimated to affect up to 22% of TB patients with cavities and 4% in those without cavities [
3]. However, in the current population, pulmonary infiltrates were the most common chest x-ray abnormality both at baseline (91%) and week 24 (41%). Only 5% of our participants had pulmonary cavities on chest x-ray at week 24. This did not change much from the 6% cavities registered at baseline. All the five patients with cavities at week 24 had a good treatment outcome (cured) but 3/5 had been diagnosed with smear negative TB at baseline. This result supports previous work indicating persistence of pulmonary cavities after successful pulmonary TB treatment [
16].
At week 24, 95% (95/100) of the chest x-rays showed marked improvement from the baseline chest x-rays and only 52% (52/100) were abnormal. Using ImmunoCAP, all patients with normal and abnormal chest x-rays at week 24 were negative for Aspergillus-specific IgG (range: <2 to 9). However, using Immulite, there were five positives among those with normal chest x-ray (median 7 [IQR: 5.9, 9.2]) and three positives among those with abnormal chest x-rays (median 7.6 [IQR: 6.3, 10.3]). We were unable to do computed tomography (CT) scans of the chest, which might have provided better definition of any residual abnormalities.
Total serum IgE is a test for general allergic disease and parasitic infections. It is commonly used together with fungal-allergy diagnostics. Using a cut-off of 170 IU/ml for total serum IgE antibodies, 71% of participants had evidence of allergic disease. This reduced to 58% (49/85) at the end of TB treatment. Of these 49 participants with elevated total serum IgE antibodies at the end of TB treatment, 12% (6/49) had elevated Aspergillus-specific IgE antibody titers in the range of 0.41 to 2.4, and Aspergillus-specific IgG antibody titers in the range of 6.7 to 55.2.
The implication of the raised total serum IgE levels in this population was not obvious. However, there is evidence that total serum IgE levels tend to be more elevated in non-asthmatic Africans than asthmatic Africans [
21]. Most scholars attribute this paradox to parasitic infestations in the African population [
21,
22]. This paradox therefore calls for the need to re-evaluate the role of total serum IgE levels in asthma in areas with a high gut parasite prevalence. Recent evidence from genotyping ancestry informative markers indicated that African ancestry is a risk factor for elevated total serum IgE levels in African admixed population [
23]. Intestinal helminths are also known to raise total serum IgE levels in HIV patients even without a fully functional CD4 T-cell repertoire [
24‐
26].
Raising the cut-off for total serum IgE to 1000 IU/mL reduced the positivity rate to more than 50%. Only 18 participants had total serum IgE greater than 1000 IU/mL at the end of TB treatment. Of these, 28% (5/18) had elevated Aspergillus-specific IgE antibody titers in the range of 0.41 to 2.4, and only two of these (2/5) had elevated Aspergillus-specific IgG antibody titers (Immulite). Long term coughing for more than three weeks at baseline was also significantly related to increased total serum IgE titers (P = 0.02). It is possible that some participants had both TB and some form of allergy at baseline; which would explain the prolonged cough durations.
53% (51/97) of the patients had negative sputum smears. Smear-negative tuberculosis was registered in 13 to 21% of our participants at baseline. Among smear negative patients (n = 51), there were 16 positive genexperts, 20 positive MGIT and 13 positive LJ cultures. Seven patients were negative for all the TB diagnostics (i.e. smear, genexpert, MGIT, LJ) but with abnormal chest x-rays. These seven patients were negative for Aspergillus IgG antibodies with very low titers at both baseline and week 24. CPA associated with TB constitutes a significant unrecognized public health problem, which is probably being incorrectly identified as ‘smear-negative tuberculosis’ especially in Africa.
Based on data from India, reduced pulmonary function is associated with
Aspergillus sensitisation, which unfortunately we were unable to measure. Post-tuberculous sequelae include CPA, bronchiectasis [
10] and to this we add
Aspergillus sensitisation. Reduced pulmonary function persists in patients cured of TB [
27], and
Aspergillus sensitisation could be one of the explanations, through poorly understood mechanisms. The prevalence of pulmonary Aspergillosis in HIV patients is underestimated in Africa because of difficulty in accurate diagnosis. We have only had a few epidemiological studies in Uganda [
4,
6,
8]. However, in collaboration with The Global Action Fund for Fungal Infections (GAFFI) and the University of Manchester, efforts are being put in place to train more mycologist and build laboratory capacity in Uganda. We hope that this might solve the problem in the near future.
Study limitations
The major limitation to the study was that we failed to access control samples for comparison of these antibody titers. This would possibly give more useful information. So we agreed to move on without the controls since the primary goal of this study was not to define diagnostic cut-offs for assays that might subsequently be used in a prevalence study in Uganda. Due to the limited published data on fungal disease epidemiology in sub-Saharan Africa, we found a major challenge in defining which diagnostic cut-offs to use for Aspergillus-specific IgG antibodies in both ImmunoCAP and Immulite. Previously published cut-offs range in 10 to 50 mg/l in similar populations. We were unable to do computed tomography (CT) scans of the chest, which might have provided better definition of any residual abnormalities. Similarly, we did not add other tests like culture since we used stored serum samples. Intestinal helminths are known to raise total serum IgE levels in HIV patients even without a fully functional CD4 T-cell repertoire [
24‐
26]. So the implication of the raised total serum IgE levels in this population was not obvious. We observed a discrepancy in the
Aspergillus-specific IgG antibody levels between the ImmunoCAP and Immulite similar to what has been described before [
15]. Cavities are known to be the major predisposing factor to CPA in TB. However, in our population, cavities were seen in only 5% of the participants.